The Senators won 3-0 despite being thoroughly outshot in expected goals: 2.55 to 3.62. This was a heist.

The Islanders controlled 59% of the xG at 5v5, took 23 shot attempts on goal to Ottawa's 16, and generated eight high-danger chances to the Senators' two. None of this mattered. L. Ullmark stopped all 23 shots he faced—a perfect 1.000 SV%—and saved 3.62 goals above what the xG model predicted. The Islanders underperformed their expected goals by 3.62, unable to convert a single high-danger chance (0 for 8).

I. Sorokin kept New York in the game (13/15 saves, 0.867 SV%), but the problem wasn't goaltending. It was Ottawa's efficiency.

The Senators were lethal on the rush, scoring two goals on just 15 such chances. When play tightened, the Islanders generated four rebound opportunities but capitalized on zero. Ottawa created three and finished one.

Michael Amadio scored on a wrist shot from the doorstep at 17:29 of the third period, putting the game away. In a matchup where New York controlled play at even strength, every marginal advantage broke Ottawa's way.

The models can't pick this game. Ottawa projects at 2.7 goals to New York's 2.4, giving the Senators 55% win probability—essentially a toss-up.

Ottawa owns the underlying play. The Senators generate 56.0% of expected goals at 5v5, compared to the Islanders' 48.0%, and control shot volume at 53.0% Corsi. This is a territorial mismatch: Ottawa should spend more time in the offensive zone. The Senators are 6-3-1 over their last 10 games, while the Islanders sit at -8 goal differential in the same stretch—a gap that suggests New York's defense has been breaking down, not just absorbing variance.

The season-long consistency matters. Both teams sport a 0.993 PDO, meaning neither can claim to be over- or underperforming their underlying numbers. Ottawa's Corsi and xG edges aren't noise—they're sustained.

Where does the Islanders' path live? In individual performances and variance. Anders Lee (1.20 xG/60, 9 goals, 20 assists, 2 high-danger goals) generates chances at a rate barely lower than Brady Tkachuk (1.19 xG/60, 11 goals, 20 assists, 2 high-danger goals), and at 6.4% shooting, Lee has room to regress upward. Bo Horvat has quietly put up 18 goals with a 12.8% shooting percentage despite logging only 0.87 xG/60—efficiency that could carry into tonight.

Ottawa's advantage flows through the depth chart. Dylan Cozens (1.00 xG/60, 13 goals, 15 assists, 1 high-danger goal) operates at a 9.5% shooting percentage, suggesting sustainable production. Tkachuk carries Ottawa's offensive load at 7.5% shooting; with higher volume, he's the likely difference-maker if play tightens.

The defensive weak links offer asymmetric risk. Nick Jensen anchors Ottawa's defense, while Alexander Romanov covers that role for New York. Romanov's likely assignment—either Tkachuk or Cozens—is where the matchup tilts. If either defender has an off night, the goal differential swings sharply.

Ottawa's territorial advantage should produce more quality opportunities. The Islanders can win if Romanov contains Tkachuk and Cozens, and if Lee and Horvat maintain their shooting efficiency. The 55% projection captures this precisely: Ottawa is favored, but it's genuinely a flip.